House of Commons - Health - Minutes of Evidence

Select Committee on Health Minutes of Evidence

Memorandum by British American Tobacco

THE TOBACCO INDUSTRY AND THE HEALTH RISKS OF SMOKING (TB 28)

BRITISH AMERICAN TOBACCO'S OWN RESEARCH

85. For more than 40 years, the operating companies of British American Tobacco (ie those companies below the holding companies) have carried out a significant amount of internal product research, aimed at exploring modifications and innovations that might produce "safer" cigarettes. Research in different countries was directed locally, pursuing different lines of enquiry and responding to different local conditions, although there was a central fundamental research facility in England at Southampton, which was funded jointly by the operating companies.

86. To our knowledge, the UK Government, and its scientific advisers, have not, with the possible exception of cigarettes lower in tar yield, determined the characteristics of a cigarette which they would call "safer". British American Tobacco agrees that the science in this area is inconclusive and, for this reason, has refrained from making health claims about any of its products. The research, however, has been guided by what governments and scientific and medical authorities have suggested as strategies for a potentially "safer" product, as well as by our own monitoring and appraisal of the scientific literature. A "safer" cigarette would, presumably, be a cigarette, the consumption of which leads ultimately to a reduction in the health risks associated with smoking; what the UK's Independent Scientific Committee on Smoking and Health ("ISCSH") called a "lower risk" cigarette (ISCSH, Second Report, p 1, 1979). Such a determination can be made only by long-term epidemiological studies, the results of which could only be known many years after the decision had been taken to place the product on the market. A great deal of uncertainty still exists today about what the practical, product-based response of tobacco manufacturers should be to the risks associated with smoking. The successes and failures of British American Tobacco's research programme, over the years, can properly be understood only in the light of the persistent absence of an answer to this fundamental question.

87. In 1956, a group research centre was established at the Southampton factory site to investigate fundamental aspects of British American Tobacco's products. A Scientific Control Laboratory had previously existed at British American Tobacco's Liverpool factory, but it had dealt mainly with processing and product quality issues. In the light of the scientific reports of an association between smoking and disease, notably lung cancer, which appeared in the early 1950s, as well as a perceived need to expand and modernise British American Tobacco's research and development capabilities, the decision to create such a centre had been taken in 1955. At the same time, British American Tobacco's Board recruited a senior scientist to act as their Scientific Adviser. Sir Charles Drummond Ellis, FRS, was primarily a mathematician and physicist, and had been a research assistant to Lord Rutherford, and, during the War years, a scientific adviser to the Chiefs of Staff (for which service he received his knighthood). Before joining British American Tobacco, he had been a scientific member of the National Coal Board. Sir Charles's eminence as a scientist gave him a key role in steering British American Tobacco's smoking and health research in its early days.

88. Despite the existing published literature on possible relationships between smoking and disease, tobacco science in the 1950s was, in many respects, in its infancy. Chemists recruited to the new Southampton research centre found themselves obliged to devise and construct equipment and methodology to conduct the smoke chemistry analyses which the research programme would require. Some of the earliest archival documents from this period refer to the work of the smoke group, which set out to analyse tobacco smoke for 3,4 benzo(a)pyrene. Another section to be created at the research centre, almost at the outset, was the filters group, charged with looking at the general and selective filtration properties of a variety of materials. The early initiatives laid the foundation for a substantial internal and external research effort which continues today. For reasons of length, it is not possible in this memorandum to do more than highlight some of the main areas of research activity.

89. This research effort has encompassed the following main areas:

— the development of a cigarette with lower biological activity;

— the selective removal or reduction of specific constituents of cigarette smoke;

— the development of alternative and innovative products; and

— the development of low tar products.

Lower biological activity cigarettes

90. British American Tobacco has carried out research into the development and evaluation of potential biological test systems and their application and use for the development of cigarettes with lower biological activity. Activity was typically measured on a per cigarette basis, or per unit weight of tar ie "specific activity". Biological activity was assessed in a range of tests involving the exposure of cells or laboratory rodents to condensate or whole smoke. Exposure conditions were generally high in order to maximise the chance of inducing quantifiable changes. This was so as to allow comparison to be made of products of different design in terms of their effect on biological activity.

91. Whilst a range of biological assays were used, three stand out as the most important. They were mouse skin-painting, inhalation studies and the Ames test. All three have been shown to respond to exposure to tar or whole smoke and to produce changes that allow comparison of products with different design parameters.

92. Janus (mouse skin-painting): One of British American Tobacco's most important early initiatives in the biological testing arena was the launch, in 1965, of a mouse skin-painting programme known as Project Janus. At this time mouse skin-painting was becoming a common place biological assay for many substances. Since 1962, of course, the company had participated in the industry mouse skin-painting programme at the Harrogate laboratories. The parallel Janus programme was devised, using very much the same experimental techniques as the TRC programme, in order that British American Tobacco could undertake prompt and confidential testing (ie confidential in terms of maintaining trade secrets) of the comparative mouse skin activity of product modifications of relevance to its own business.

93. Project Janus ran at the Battelle contract laboratories in Frankfurt, Germany, for some 14 years, during which time 15 experiments were conducted. Project Janus looked, for example, at the effects on mouse skin activity of certain tobacco extraction and reconstitution processes, and at the biological activity of certain commercially available, chemically treated, tobacco sheets. As at Harrogate, it was found that most of the modifications tested had relatively small effects on specific activity. The use of some of the sheet materials did have a more marked effect on activity, but there seemed no prospect of incorporating significant levels of such sheets in a cigarette acceptable to consumers.

94. Inhalation studies: An important focus of the scientific literature, during the 1970s, was the need to develop an inhalation assay, by which the respiratory tracts of laboratory rodents could be exposed to fresh, whole smoke. Despite the considerable physiological differences between the respiratory systems of humans and test animals, it was thought that such an assay would most closely approximate the conditions of human smoking.

95. Battelle undertook some exploratory inhalation testing, under contract to British American Tobacco, in 1971 and 1972. By 1974, however, a Life Sciences Division had been established at the Southampton research centre, equipped to conduct inhalation testing on laboratory rodents. This Division was also concerned to put comparative toxicity testing of tobacco products, using an inhalation bio-assay, on a sound scientific footing. Significant success was achieved in methodological research, which led to the publication of a series of papers describing British American Tobacco's short-term inhalation exposure regimes. Much of the work of the Life Sciences Division focused on the evaluation of different prototypes of British American Tobacco's proprietary tobacco substitute material, BATFLAKE. This was developed in line with the recommendations of the ISCSH and tested under its protocol.

96. These short-term assays could assess the comparative toxicity of the smoke from different cigarettes, at a number of defined sites in the rat respiratory tract. The effects reported, including keratinisation, metaplasia and hyperplasia, rapidly regressed following cessation of smoke exposure. These were not necessarily pre-cancerous changes and, indeed, the short-term assays were not used as tests of carcinogenicity.

97. Scientists at Southampton conducted one pilot long-term inhalation assay, which they hoped might elucidate the carcinogenic potential of whole, fresh smoke in such a test, but the results of this study are consistent with the conclusion expressed by the US Surgeon General in his 1982 Report: "Attempts to induce significant numbers of bronchogenic carcinoma in laboratory animals were negative in spite of major efforts with several species and strains" (US Surgeon General, "The Health Consequences of Smoking: Cancer", p218, 1982).

98. The Ames test: A further important biological research effort undertaken by British American Tobacco in the 1980s was the attempt to develop a cigarette with reduced mutagenic activity, as measured by a widely used in vitro assay, the Ames test (names after its inventor, Professor Bruce Ames).

99. Professor Ames had developed his test in 1975, as a relatively sensitive test for the potency of chemicals to cause mutations in the DNA of the bacterium Salmonella typhimurium. The test, which is comparatively rapid, does not use animals and was, therefore, of great interest to ourselves as it was to the chemical and pharmaceutical industries.

100. There are particular problems in applying the Ames test to tobacco smoke condensate because of its cellular toxicity and complex chemical composition. In addition, much of the initial interest in the test was prompted by the historic belief that most mutagens were also potential carcinogens. In hindsight this has been shown not to be the case. Some reported carcinogens test negatively for mutagenicity, using Ames, while other substances which are not believed to be carcinogens give a positive response. This assay has some utility in certain and defined circumstances.

101. Significantly for British American Tobacco's biological research effort, comparative toxicity rankings of cigarette design variables using the Ames test do not always coincide with the rankings derived from mouse skin-painting. For example, Burley tobacco is more active than flue-cured tobacco in the Ames test, but less active in a mouse skin-painting assay. Indeed, the pattern of inconsistent rankings can be found across the range of biological tests which British American Tobacco has employed to assess its products. These observations clearly indicate the inherent difficulty and need for significant scientific judgements when interpreting different responses in a range of biological test systems. Nevertheless, British American Tobacco continues to this day to explore the potential use of biological tests in assessing product modifications.

Selective Reduction

102. Some of the earliest research at the Southampton research centre involved identifying and measuring constituents of tobacco smoke and, in the case of constituents reported to present a possible hazard, searching for ways selectively to remove them or to reduce their yield. "Selective" reduction, as the name implies, is defined as achieving reduced yield of target compounds against some constant index of overall yield. For example, reducing tar yield will reduce the yield of each of the chemical constituents of tar. Selective reduction of one of those constituents implies reduction against constant tar yield.

103. Smoke created by burning organic material of any kind is chemically complex. It is widely reported that there are more than 4,000 chemical constituents of cigarette smoke, most of them present at only trace quantities. Doubtless, the ongoing refinement of analytical techniques will identify further components present at even lower levels. Over the years a significant number of the compounds found in tobacco smoke have been hypothesised as contributing to smoking related diseases.

104. British American Tobacco's research has examined a wide variety of individual compounds and groups of compounds, including phenols, aldehydes, nitrosamines, poly-aromatic hydrocarbons (including 3,4benzo(a)pyrene), oxides of nitrogen and carbon monoxide.

105. The selective reduction of some target compounds can be achieved by filtration, altering the blend, amending cultivation methods and other techniques, but rarely, if ever, without consequential effects for the rest of the smoke. Furthermore, selective reduction of constituents does not necessarily lead to clear and consistent reductions in activity in biological tests.

106. Although a great deal of work has been done on selective reduction research, it has not led to the development of a commercially acceptable cigarette, which might be recognised as "safer", for a number of reasons:

— There are many smoke constituents which have been suggested as being responsible for diseases in smokers, making the benefits of reducing any particular constituent or group of constituents uncertain;

— There are immense practical problems in trying to reduce one constituent or a group of constituents;

— Some of the more innovative proposals to modify the product have involved adding substances or compounds which, upon further investigation, were themselves found to transfer to the smoke. In some instances, there have been questions raised as to whether such substances or compounds might themselves have adverse health implications;

— Sometimes the removal of smoke constituents results in an increase in biological activity;

— Some design modifications, especially the use of novel filters, have an adverse effect on the taste of the smoke, resulting in a product that would be unacceptable to smokers.

107. Despite internal and external investigation, the significance of particular constituents at the levels found in smoke is unknown. In the end, therefore, it was found that the overall reduction of smoke yields, as well as being more logical in scientific terms, was the simpler and more practical way to reduce the deliveries of individual constituents. This approach is wholly consistent with the outcome of the biological test programme described previously. By using filters and ventilation to reduce tar, it is possible to reduce the level of most smoke constituents more or less to the same extent as tar. Thus, in our lower tar products, there has been a reduction in the levels of most of the constituents which have been of concern to governments and medical bodies.

Alternative and innovative products

108. Over the years, a third important strand of British American Tobacco's own smoking and health research has been to explore the design of some novel products which might, on the basis of claims made about smoking and health, be regarded as "safer". To date, none of these products has been successful. The impetus for such research came, in the late 1950s and 1960s, from Sir Charles Ellis himself. Sir Charles clearly believed, not only that nicotine played an important role in smoking motivation, but that it was beneficial in helping smokers to cope with the stresses and strains of everyday life. At the same time, he took seriously the hypothesis that the products of tobacco combustion might play a causal role in human disease. His solution was to direct the development of an alternative smoking article, which delivered nicotine without the products of tobacco combustion. Internally, the article was known as Ariel.

109. Several Ariel prototypes were constructed and tested, but in essence the Ariel concept had an elongated body, with a mouthpiece end and an outer end, simulating a cigarette. Enclosed within this article was an annular tube with nicotine. Burning the tobacco in a conventional manner caused the nicotine in the tube to heat and vapourise. As a result the smoker would receive an aerosol, very largely composed of nicotine without any of the pyrolytic or combustion products normally associated with cigarette smoke.

110. Among the research challenges presented were controlling the heating mechanism, finding a suitable carrier for the nicotine, presenting a palatable vapour to the smoker, and guarding against high and aversive nicotine yields. An equally complex development problem was to find a way of manufacturing the article at the volume and cost that market acceptance would require.

111. In the event, these problems could not be overcome, and after several years of research it was recognised that, while the Ariel concept was in principle feasible, the development had been a technical failure. An Ariel prototype was patented, but no commercial model has been produced. The results from smoking panel tests showed that Ariel was unacceptable to smokers.

112. British American Tobacco continued to look at the possibility of developing alternative smoking articles, but not until the 1980s was further systematic research conducted. At around the same time R J Reynolds in the United States was preparing to market Premier, a product similar in some respects to the original Ariel concept. British American Tobacco attempted to develop an article which would heat, rather than burn, tobacco, and which would have other features such as reduced sidestream smoke. Again, there was a failure to resolve the technical problems created by the product specification. The Reynolds product proved to be unacceptable to consumers when launched, and it was clear that British American Tobacco's own internal development would have faced similar acceptability problems, even if the practical obstacles to its development could have been overcome.

113. Premier also ran into regulatory problems. In 1988 the American Medical Association and others petitioned the FDA to regulate the new product as a drug, specifically a nicotine delivery device, which thereby would not be as freely available as cigarettes. US Surgeon General C Everett Koop also pressed for FDA regulation. The regulatory controversy was not resolved by the time Premier was withdrawn from the market due to lack of consumer acceptance. The fate of Premier shows just how hard it is to develop an alternative smoking article, or other innovative product, which is acceptable to both consumers and the regulatory authorities.

114. At the end of the 1980s, British American Tobacco embarked on the search for a conventional cigarette with ultra-low tar yield, and a greatly modified tar to nicotine ratio. This project became known internally as Greendot. The project was responsive to the direction suggested by the ISCSH, and the ISCSH were appraised of the project. The Greendot targets were very ambitious, and the project encountered the familiar problems of practicability and acceptability.

115. Neither the alternative smoking article approach to a potentially lower risk cigarette, nor the highly modified conventional cigarette approach, have yet borne fruit as commercially available products. These are, however, areas of ongoing research.

Lower tar products

116. Since at least the 1950s, when there were widely publicised reports that painting tobacco smoke condensate on to the shaved backs of mice caused the development of tumours (Wynder E L and Graham E A, "Tobacco Smoking as a Possible Etiologic Factor in Bronchiogenic Carcinoma", JAMA, 143 (4):336, 1950), scientists have argued that the particulate matter in cigarette smoke might be responsible for causing cancer in smokers. This particulate matter (as opposed to the gaseous components of smoke) is sometimes known as "tar", although, strictly speaking, "tar" is usually defined as particulate matter with the water and nicotine extracted.

117. An important part of British American Tobacco's research activity during the last forty years has been to produce products with lower tar yields as tested on standard smoking machines. To date this has been the only product modification that has met with largely consistent support from governments. This simpler and more practical option has not only seemed to be common sense (given the reported dose response relationship between smoking and lung cancer), but also has the effect in most cases of unselectively reducing yields of many of the specific constituents targeted as potential health hazards. It should also follow from the reported dose response relationship that reducing the total amount of smoke will reduce biological activity, and there are experimental data to support this proposition. Moreover, it has been possible to develop commercially acceptable low tar products without the practical problems encountered with the other strands of "safer cigarette" research ie lower biological activity cigarettes, selective reduction and alternative or innovative products.